Using ROC Curves to Choose between Competing Classification Schemes

A bank is interested in predicting whether or not customers will default on loans. You have proposed a model that is based on a subset of the available predictors, and you now need to show that its results are better than those from a simpler model currently in use and no worse than results from a more complex model.

The model results are collected in bankloan.sav. See the topic Sample Files for more information. Use ROC Curve to compare the predictive abilities of these three models.

Next